Abstract
We propose a hybrid representation of syntactic structures, combining constituency and dependency information. The headed constituency trees that we use offer the advantages of both those approaches to representing syntactic relations within a sentence, with a focus on consistency between them. Based on this representation, we introduce a new constituency parsing technique capable of handling discontinuous structures. The presented approach is centred around head paths in the constituency tree that we refer to as spines and the attachments between them. Our architecture leverages a dependency parser and a large BERT model and achieves 95.96% F1 score on a dataset where \(\approx \)10% of trees contain discontinuities.
Work supported by POIR.04.02.00-00-D006/20-00 national grant (Digital Research Infrastructure for the Arts and Humanities DARIAH-PL).
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
We use examples in Polish, since its system of 7 grammatical cases makes the grammatical relations more easily visible than in English.
- 2.
- 3.
- 4.
See http://git.nlp.ipipan.waw.pl/constituency/spines-attachments for the code and dataset.
- 5.
In other words, the tasks (1) and (2b) can be seen as converting the dependency structure to constituencies. Note, however, that the constituency trees are more detailed, so this process adds information.
- 6.
We use the lower subscript NP \(_i\) to differentiate between two different NP nodes, and not to introduce a separate category NP \(_i\).
- 7.
- 8.
- 9.
- 10.
Since the number of distinct spines is fairly limited, we decided to treat them as atomic labels.
- 11.
For the bracketings metric, each span is counted only one time, e.g. for the tree in Fig. 4, (Dam) and (Dam) are treated as the same span (Dam) etc.
- 12.
We noticed that when validation data loss was used for early stopping and model selection, the accuracies on validation data still exhibited a growing tendency.
- 13.
Let \(TP_l\), \(FP_l\), \(FN_l\) denote the number of true positives, false positives and false negatives respectively for label l in evaluation data. For a set of labels S, we calculate the aggregate precision P\(_S\) as \(({\sum _{l \in S}TP_l})/({\sum _{l \in S}TP_l + FP_l})\), i.e. the proportion of correctly predicted labels from S to all predicted labels from S. The aggregate recall R\(_S\) is \(({\sum _{l \in S}TP_l})/({\sum _{l \in S}TP_l + FN_l})\), i.e. the proportion of correctly predicted labels from S to all gold labels from S. The aggregate F1\(_S\) is the harmonic mean of P\(_S\) and R\(_S\).
- 14.
References
Brants, S., et al.: TIGER: linguistic interpretation of a German corpus. J. Lang. Comput. 2, 597–620 (2004). https://doi.org/10.1007/s11168-004-7431-3
Coavoux, M., Cohen, S.B.: Discontinuous constituency parsing with a stack-free transition system and a dynamic oracle. In: Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pp. 204–217. Association for Computational Linguistics, Minneapolis, Minnesota (2019). https://doi.org/10.18653/v1/N19-1018
Fernández-González, D., Martins, A.F.: Parsing as reduction. In: Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing (Volume 1: Long Papers), Beijing, China, pp. 1523–1533. Association for Computational Linguistics (2015). https://doi.org/10.3115/v1/P15-1147
Gerdes, K., Guillaume, B., Kahane, S., Perrier, G.: SUD or surface-syntactic universal dependencies: an annotation scheme near-isomorphic to UD. In: Universal Dependencies Workshop 2018, Brussels, Belgium (2018). https://hal.inria.fr/hal-01930614
Gómez-Rodríguez, C., Vilares, D.: Constituent parsing as sequence labeling. In: Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, Brussels, Belgium, pp. 1314–1324. Association for Computational Linguistics (2018). https://doi.org/10.18653/v1/D18-1162
Kitaev, N., Cao, S., Klein, D.: Multilingual constituency parsing with self-attention and pre-training. In: Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, Florence, Italy, pp. 3499–3505. Association for Computational Linguistics (2019). https://doi.org/10.18653/v1/P19-1340
Klimaszewski, M., Wróblewska, A.: COMBO: State-of-the-art morphosyntactic analysis. In: Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing: System Demonstrations, Punta Cana, Dominican Republic, pp. 50–62. Association for Computational Linguistics (2021). https://aclanthology.org/2021.emnlp-demo.7
Marcus, M.P., Santorini, B., Marcinkiewicz, M.A.: Building a large annotated corpus of English: the Penn Treebank. Comput. Linguist. 19(2), 313–330 (1993). https://aclanthology.org/J93-2004
Mroczkowski, R., Rybak, P., Wróblewska, A., Gawlik, I.: HerBERT: Efficiently pretrained transformer-based language model for Polish. In: Proceedings of the 8th Workshop on Balto-Slavic Natural Language Processing, Kiyv, Ukraine, pp. 1–10. Association for Computational Linguistics 2021. https://www.aclweb.org/anthology/2021.bsnlp-1.1
Nivre, J., et al.: Universal dependencies v2: An evergrowing multilingual treebank collection. CoRR, abs/2004.10643 (2020)
Seddah, D., et al.: Overview of the SPMRL 2013 shared task: a cross-framework evaluation of parsing morphologically rich languages. In: Proceedings of the Fourth Workshop on Statistical Parsing of Morphologically-Rich Languages, Seattle, Washington, USA, pp. 146–182. Association for Computational Linguistics (2013). https://aclanthology.org/W13-4917
Świdziński, M., Woliński, M.: Towards a bank of constituent parse trees for Polish. In: Sojka, P., Horák, A., Kopeček, I., Pala, K. (eds.) TSD 2010. LNCS (LNAI), vol. 6231, pp. 197–204. Springer, Heidelberg (2010). https://doi.org/10.1007/978-3-642-15760-8_26
Woliński, M.: Automatyczna analiza składnikowa języka polskiego. Wydawnictwa Uniwersytetu Warszawskiego, Warsaw (2019). https://www.wuw.pl/data/include/cms/Automatyczna_analiza_skladnikowa_Wolinski_Marcin_2019.pdf
Woliński, M., Hajnicz, E.: Składnica: a constituency treebank of Polish harmonised with the Walenty valency dictionary. Lang. Res. Eval. 55(1), 209–239 (2021). https://doi.org/10.1007/s10579-020-09511-7
Wróblewska, A.: Polish Dependency Parser Trained on an Automatically Induced Dependency Bank. Ph.D. dissertation, Institute of Computer Science, Polish Academy of Sciences, Warsaw (2014). http://nlp.ipipan.waw.pl/Bib/wro:14.pdf
Zhou, J., Li, Z., Zhao, H.: Parsing all: syntax and semantics, dependencies and spans. In: Findings of the Association for Computational Linguistics: EMNLP 2020, pp. 4438–4449. Association for Computational Linguistics (2020). https://doi.org/10.18653/v1/2020.findings-emnlp.398
Zhou, J., Zhao, H.: Head-driven Phrase Structure Grammar parsing on Penn Treebank. In: Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, Florence, Italy, pp. 2396–2408. Association for Computational Linguistics (2019). https://doi.org/10.18653/v1/P19-1230
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Krasnowska-Kieraś, K., Woliński, M. (2023). Constituency Parsing with Spines and Attachments. In: Mikyška, J., de Mulatier, C., Paszynski, M., Krzhizhanovskaya, V.V., Dongarra, J.J., Sloot, P.M. (eds) Computational Science – ICCS 2023. ICCS 2023. Lecture Notes in Computer Science, vol 14073. Springer, Cham. https://doi.org/10.1007/978-3-031-35995-8_14
Download citation
DOI: https://doi.org/10.1007/978-3-031-35995-8_14
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-35994-1
Online ISBN: 978-3-031-35995-8
eBook Packages: Computer ScienceComputer Science (R0)